Rapid technological advances, especially in telematics and Big Data analytics, as well as the increasing penetration and use of information technology by drivers (e.g. smartphones), provide new capabilities for monitoring and analyzing driving behaviour. This paper examines the impact of driver feedback delivered through a smartphone application on driving behaviour risk indicators, within a 21-month multiphase naturalistic driving experiment involving a sample of 175 car drivers. First, a preliminary analysis utilizing summary statistics and Wilcoxon signed-rank test shed light upon the effects of upgraded feedback features on key risk indicators across experiment phases. Subsequently, Structural Equation Models (SEM) on a 73,869 trip dataset provided significant insights into how feedback mechanisms and exposure factors influence driving behaviours. Results indicate that the examined feedback mechanisms are effective in reducing the percentage of speeding time and harsh braking events, although there is an increase in harsh accelerations, which may require further refinement of the feedback system. The scorecard feature has the highest positive effect, indicating its crucial role in modifying driving habits, with gamification (competition and challenges) being the second most influential feedback mechanism. Regarding the exposure indicators, driving during the morning peak is associated with more unsafe behaviour (e.g., speeding, harsh accelerations), while driving during the afternoon peak exhibit fewer risky driving indicators (e.g., harsh braking). Additionally, results showcase strong positive correlations between speeding, harsh braking, and harsh accelerations highlighting the interconnected nature of risky driving behaviours. These findings may be beneficial for insurance companies, fleet management applications, and policymakers, enabling them to leverage results to improve traffic safety and driver behaviour.
| ID | pj289 |
| DOI | |
| Tags |
